London NorthWestern Railway is a sub-brand of the West Midlands Trains franchise. It runs most of the stopping and semi-fast services along the southern and Midlands section West Coast Main Line. In May 2019, there was a major timetable change. These changes resulted in a significant drop in reliability of LNWR services. Using these timetable changes as a case study, this article explores why complex and asymmetrical timetables on very congested networks simply don't work, unless there is a significant amount of slack/recovery time in the schedule.
Previously, there were a number of services from London (3 trains per hour (tph)), Liverpool (2tph) and Rugeley (1tph) which terminated at Birmingham New Street. With the electrification of the Chase Line (Birmingham to Rugeley) and increase in demand for those services, a second train per hour was required between the two. Additionally, the semi-fast Trent Valley services needed to be diverted between Stafford and Crewe. Instead of running via Stoke-on-Trent, they would run direct. Therefore, an additional service between Birmingham and Crewe was introduced to pick up the stations no longer served by the London to Crewe services. In response to the additional services, it was decided to combine the South to Birmingham and Birmingham to North services into several South to North via Birmingham services.
But, because the through services are a combination of existing ones, the end-to-end journeys are asymmetrical. 3 trains per hour run from London, with a 4th from Birmingham International. From New Street, they run to Liverpool (2tph), Rugeley (2tph) and Crewe (1tph). One service per hour to/from London serves two destinations north, whilst all the others serve one. Northbound they run as follows:
Even within the services to London, the stopping pattern is not homogenous. They run as follows:
* = Stops only going towards Birmingham
^ = Stops only going towards London
With that variation, even the 1tph London to Liverpool, Crewe and Rugeley services become asymmetrical. Taking the variation of stopping pattern into account, the services look like this:
In fact, the only service that is actually the same in both directions is London (3) to Liverpool, but even then one has to remember the variation in stopping pattern just in the different London, and the fact that, in the southbound direction, it has to attach to a service from Crewe.
And, if you haven't been able to take all of that in, or have had to spend an awful lot of time digesting the information before understanding, then that's the point. It's too complicated.
The diversion of the Trent Valley semi-fast service was because there had been a significant increase in demand meaning that running 4 car trains all day was not an option. Because of the level crossing at Stone station, only 4 car trains can stop there. Thus, that was limit on the number of coaches that could be operated on the Trent Valley services. By changing the stopping pattern and running directly between Stafford and Crewe, longer trains could run. This method was already used at peak times from London before the timetable change.
However, the additional Crewe and Rugeley to Birmingham trains meant that there would be too many services terminating at Birmingham New Street. As I explored briefly in my first HS2 article, Birmingham New Street has too many trains using it, and too few platforms to accommodate them all. Through (non-terminating) services occupy platforms for a shorter amount of time, meaning that more trains can be accommodated if more trains run through rather than terminating. Building more platforms is simply not an option because the station is located directly in the city centre with no space around it to make the station larger.
The very complicated timetables mean that delays spread and multiply much more quickly. Previously, rolling stock was mostly self-contained. If a Liverpool to Birmingham train was late, that would effect other trains which came into contact with it (meant from a timetabling point of view. IE: the delay to the Liverpool to Birmingham service would possibly cause other trains around it to be delayed) and the services that that train and crew would drive next. But, if this happens in the new timetable, this means that the service actually delayed is Liverpool to London. This significantly increases the number of “contact trains” which could be delayed. The longer services also mean that there is more of a potential for that train to pick up delays.
Therefore, there should be more slack built into the timetable to combat potential delays. A look at the London to Rugeley services sees 6 minutes of engineering allowances (which must be included in order to make the timetable compliant). There isn't much scope to make up time within station dwell times either. Bar a 3 minute wait at Northampton, and a 4-7 wait at Birmingham New Street, there is no other location with any significant dwell time. With a journey time of over 3 hours, and a network as congested as this one is, the maximum possible 9 minutes of time that could theoretically be made up (I say 'theoretically' because engineering allowances also allow for the variations in platform positions in a large station that trains use, thus meaning that not all of these are minutes that a driver can actually make up) is not very much at all. It's certainly nowhere near enough to get to the Swiss level of punctuality (a level only maintained because of generous timetabling allowances).
In the most recent statistics (Aug to Sep 2019), 72.9% of LNR trains ran within 5 minutes of their scheduled time, but only 40.2% actually ran on time. 6.8% of trains were cancelled. The London – Midlands – North West sector (the one with the complex timetables) is worse. 63.5% of trains ran within 5 minutes, and 27.7% ran on-time. 10.1% of were cancelled.
This is seriously bad. Given that the national average is 90.1% within 5 minutes, 64.1% on time, and 4.1% cancelled, LNWR are performing awfully.
Continuing with the London to Rugeley services, 33 are supposed to run every day. An average of 4 are cancelled. Worse, the most likely train to be cancelled is the 17:23 (bang in rush hour), running 68% of the time since the 17th of June (the earliest date for which statistics are available). The Chase Line from Birmingham to Rugeley via Walsall has arguable been the worse hit. The contingency plan for a delayed train from London to Rugeley is to turn it back at Hednesford (between Walsall and Rugeley) so that it is on time for its return journey. While this is sensible, the complexity of the timetables means that trains regularly go over the threshold meaning that trains frequently turn back there, leaving Rugeley passengers hanging around at Hednesford for hours (not an exaggeration). Especially when other parts of the service are disrupted, the Chase Line is the first to be sacrificed. For example, the most recent serious disruption was a track circuit failure between Long Buckby and Rugby. There were no trains from Birmingham to Rugeley between 14:20 and 18:53. While this is an extreme example, it demonstrates how the new timetables are being made to work. Block cancellations of 2 trains (IE: a gap of 90 minutes) is not uncommon.
Longer turnaround times at Rugeley Trent Valley could be one. The current times of 6-9 leave very little slack to make up delays from the previous trip. This is one of the reasons why trains are so regularly flipped short at Hednesford.
Originally, I had thought that one could stable a train at Rugeley as a “hot spare” (some rolling stock that could be quickly put into action should it be required). That way, if there is a delay, one just has to bring out the hot spare and run it to time back towards Birmingham and London. When the trains start arriving on time again, one puts another unit back in the stabling point. To make it work from a crew perspective, make the crew diagrams such that every crew drops one train back at Rugeley. IE: the crew arriving at XX:10 would work the XX:46 back, rather than the XX:19.
However, this has problems. For a start, Rugeley has too few facilities for it to be used as a crew resting/drop back point. Secondly, there is no suitable storage space at that end of the Chase Line for such a unit.
Capacity issues at Birmingham New Street also means that it is not possible to go back to the previous situation where trains from London, Rugeley, Crewe and Liverpool terminated there.
In fact, the more one looks, the more one realises that it finding solutions is not possible. There always appears to be one factor that messes up the plan.
Professional timetablers and planners will have to find some sort of solution, because this poor service cannot continue. But, I do not envy the people who have to find said solution.